Scottish Government’s reconviction rate is calculated as the average number of reconvictions per offender (and was previously a national indicator). Not all demographic groups have seen this fall in this measure, however. For example, the reconviction rate for women over the age of 30 has increased, despite the overall fall in the reconviction rate.
Figure Two here - graph of these
We’ve also seen substantial changes over time in number of people convicted (known as the prevalence of conviction)2 Ben wrote about these long-term differences in conviction rates in Scotland by age and sex with Jon Minton here., again with divergent trends for men and women of different ages. These trends in prevalence of conviction have a knock-on effect on the reconviction rate because they affect the proportions of people of different ages who are convicted in a given year. These proportions - the ‘offender mix’ - can influence the overall reconviction rate even if the reconviction rates don’t change for any age group. This is because different age and sex groups typically have different reconviction rates3 Typically young people will be convicted at higher rates than older people, and men at higher rates than women., and the overall reconviction rate the is just average of the conviction rates for each group, weighted by the size of the group. As a result, changes in the proportions of younger to older people and men to women convicted can affect the overall reconviction rate without groups’ reconviction rates changing at all. This confusing phenomenon is known as Simpson’s paradox, and has been identified in places as diverse as statistics on kidney-stone removals to baseball batting averages.
One implication of this paradox is that the fall in the reconviction rate is not only a measure of changes in reoffending - it also reflects changes in the mix of people coming into the justice system (i.e. being convicted) in the first place. This also means that we don’t know exactly how much of the fall in the overall reconviction rates is due to reductions in reoffending, and much is due to these changes in offender mix, which may reflect things like diversion from prosecution or social changes which mean that fewer people offend.
What can we do about it?
This problem - disentangling the effects of change in offender mix from other factors affecting overall population rates of some variable we’re interested in - is a common one in demography, where researchers want to understand how much differences between rates (such as mortality rates) in different countries are due to their different population structures.
We can use the methods demographers have developed to apportion the amount of change in the reconviction rate that is due to actual reductions in reoffending, and how much is due to changes in the mix of people convicted4 The code and data used for our analysis is available here. Specifically, we use the pair of techniques known as standardization and decomposition.
Standardization shows us what the reconviction rate would have been under different scenarios - for example, if there was no change in the prevalence of conviction for different demographic groups (no change in the offender mix) or if there was no change in the frequency of conviction for different demographic groups - and decomposition gives us the percentage of the difference in rates between two years attributable to each of the factors we have included in the standardization.
Results
Following a method developed by Das Gupta, we can examine the reconviction rate (the average number of reconvictions per offender) when adjusting for:
- the age/sex structure of the convicted population (the offender mix)
- the prevalence of reconviction (the proportion of offenders who reconvict)
- and the frequency of reconvictions (the average number of reconvictions per reconvicted offender)
Put in results here
may be We also examine how much of the change in prevalence was due to underlying population change in the numbers of men and women of different ages in Scotland5 Population ageing can reduce conviction rates, as young people are more likely to commit crime than older people - but only very slowly (Zimring, 2007).
What have we learned?
There are four main takeaways from these results:
First: all three factors of prevalence, frequency and age-structure have contributed to the falling reconviction rate. Probably frequency more than the other two.
Second: Our results show that men and women of different ages have very different trends in conviction and reconviction. The overall fall in the reconviction rate has been driven predominately by young people, and young men in particular. We need to recognize these age and sex differences in how convictions patterns have changed in Scotland, and not assume that the overall fall has been seen for all groups.
Third: As well as supporting people who come into contact with the justice system, we also need to broaden our understanding of the causes of this fall in prevalence, and why some groups - but not others - have become less likely to be convicted in the first place.
In this blog we’ve looked at the causes, of the change in the reconviction rate, but not the causes of the causes. In Sweden, researchers have shown that there’s a distinct socio-economic status gradient to the crime drop (20176 Note: This paper is behind a paywall.): “Those from a less well-resourced background have come to account for an increasingly large proportion of registered crime, whereas the reverse is true for those from a more affluent background.” This suggests thatwe need to understand the interaction between crime and wider inequalities in Scottish society. The Understanding Inequalities project aims to understand how inequalities in Scottish society interset, and how they imapct on crime and responses to crime. You can find out more about Understanding Inequalities’s work on crime here.
